Update README.md
Browse files
README.md
CHANGED
|
@@ -73,21 +73,6 @@ Based on evaluation with 1,000 test samples:
|
|
| 73 |
| F1 Score (Macro) | 71.90% |
|
| 74 |
| Eval Loss | 0.0834 |
|
| 75 |
|
| 76 |
-
## Training Details
|
| 77 |
-
|
| 78 |
-
### Training Configuration
|
| 79 |
-
- Learning Rate: 2e-5
|
| 80 |
-
- Batch Size: 16
|
| 81 |
-
- Number of Epochs: 3
|
| 82 |
-
- Warmup Ratio: 0.1
|
| 83 |
-
- Weight Decay: 0.01
|
| 84 |
-
- Max Length: 128
|
| 85 |
-
|
| 86 |
-
### Hardware
|
| 87 |
-
- Trained on: NVIDIA GPU
|
| 88 |
-
- Training Time: ~79 seconds
|
| 89 |
-
- Optimization: Designed for deployment on Raspberry Pi 5
|
| 90 |
-
|
| 91 |
## Intent Classes
|
| 92 |
|
| 93 |
The model can classify the following 13 intents:
|
|
@@ -180,13 +165,6 @@ session = ort.InferenceSession("model.onnx")
|
|
| 180 |
# ... inference code ...
|
| 181 |
```
|
| 182 |
|
| 183 |
-
## Dataset
|
| 184 |
-
|
| 185 |
-
This model was trained on an augmented version of the VN-SLU dataset, which includes:
|
| 186 |
-
- Original recordings from 240 Vietnamese speakers
|
| 187 |
-
- Augmented samples using various techniques
|
| 188 |
-
- Smart home specific vocabulary and commands
|
| 189 |
-
|
| 190 |
## Citation
|
| 191 |
|
| 192 |
If you use this model, please cite:
|
|
|
|
| 73 |
| F1 Score (Macro) | 71.90% |
|
| 74 |
| Eval Loss | 0.0834 |
|
| 75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
## Intent Classes
|
| 77 |
|
| 78 |
The model can classify the following 13 intents:
|
|
|
|
| 165 |
# ... inference code ...
|
| 166 |
```
|
| 167 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 168 |
## Citation
|
| 169 |
|
| 170 |
If you use this model, please cite:
|